-- twitter.lua

http = require"socket.http"
r = http.request(string.format("http://%s:%s@twitter.com/statuses/friends_timeline.json", arg[1], arg[2]))
for i,v in ipairs(json.decode(r)) do print(v.user.name .. ":", v.text, "(" .. v.created_at .. ")") end

Run like this with ‘watch’

$ watch -n 60 lua -lluarocks.require -ljson -lsocket twitter.lua <username> <password>

This snippent requires luasocket and json4lua libraries. You should get’em via luarocks

First of all let’s get nginx installed on our server, so go and get nginx tarball from nginx homepage. Download and unzip it to your home directory like this:

$ cd
$ wget http://sysoev.ru/nginx/nginx-0.6.34.tar.gz
$ tar -xzvf nginx-0.6.34.tar.gz

Enter to the unzipped directory and perform a user-wide installation with its prerequisites.

$ cd
$ apt-get install libpcre3-dev
$ apt-get install zlib1g-dev
$ mkdir nginx
$ cd nginx-0.6.34
$ ./configure --prefix=$HOME/nginx
$ make
$ make install
$ cd
$ rm nginx-0.6.34 -rf

In the end you should have a nginx directory with a structure like this

nginx/client_body_temp
nginx/conf
nginx/fastcgi_temp
nginx/html
nginx/logs
nginx/proxy_temp
nginx/sbin

If you wonder why i preferred nginx instead of httpd (Apache Http Server) or lighttpd, my answer will not be so long. Our primary goal to use a http server in the frontend of hunchentoot is to serve static files (images, videos, javascript or css files etc.) efficiently so nginx has small memory footprint compared to httpd and very stable memory management compared to lighttpd. Yes i know lighttpd is also a nice web server with very well known modules handled on an easy configuration file but it has a very annoying memory leak problem that has not been solved yet (#758). Also you should read a post about nginx vs .lighttpd comparison here. For more information please refer to the survey of web server usage statistics at netcraft because i’ll not continue to talk about web server challanges anymore. There is a probably more adequate reason for me is that i loved nginx

Anyway, you can test your nginx installation:

$ cd
$ cd nginx
$ sudo sbin/nginx
$ curl localhost

If you see the nginx response you’re so lucky that you jump over the nginx installation flawlessly

Now let’s put hunchentoot on work starting with installation of sbcl first. Go to sbcl platform download page, choose an appropriate architecture on Linux row and get it with wget.

$ cd
$ wget http://puzzle.dl.sourceforge.net/sourceforge/sbcl/sbcl-1.0.23-x86-linux-binary.tar.bz2
$ tar -xjvf sbcl-1.0.23-x86-linux-binary.tar.bz2

Installation is quite easy:

$ cd
$ mkdir sbcl
$ cd sbcl-1.0.23-x86-linux-binary
$ INSTALL_ROOT=$HOME/sbcl sh install.sh
$ cd
$ rm sbcl-1.0.23-x86-linux-binary -rf

And now set ‘SBCL_HOME’ environment variable and add ‘$HOME/sbcl/bin’ to the ‘$PATH’ environment variable to get sbcl work anywhere

$ cd
$ echo "export SBCL_HOME=$HOME/sbcl/lib/sbcl" >> .bashrc
$ echo "PATH=$PATH:$HOME/sbcl/bin" >> .bashrc
$ . .bashrc

Great! Now see the sbcl works

$ sbcl

If you entered to the sbcl interpreter you got it.

The funniest part is when you are in a lisp interpreter, now we have to install hunchentoot to the system. Here we’ll require the help of ASDF (Another System Definition Facility) and ASDF-INSTALL (Remote or tarball interface tool for ASDF). ASDF is library dependency resolver system for common lisp world. Fortunately sbcl comes with ASDF and ASDF-INSTALL.
Hunchentoot is full blown pure common lisp web server. The structure of hunchentoot is fun to digg further and plays very well while designing web frameworks. I plan to write another post to give a hunchentoot tutorial in detail. Anyway..

Now you are ready to install hunchentoot but first we should better to define a variable that skips GPG signatures of lisp libraries globally.

$ cd
$ echo "(defparameter ASDF-INSTALL-CUSTOMIZE::*VERIFY-GPG-SIGNATURES* nil)" > $HOME/.asdf-install

Install the hunchentoot by ASDF_INSTALL. This tool resolves library names from cliki, the community common lisp wiki and library repository.

* (require 'asdf-install)
* (asdf-install:install 'hunchentoot)
* (quit)

In the end you’ll got the funniest pure lisp web server in the world. Start server by typing

$ cd
$ cat > hunch.lisp
(require 'asdf)
(asdf:operate 'asdf:load-op 'hunchentoot)
(hunchentoot:start-server :port 8080)
(read)
[CTRL-D]
$ sbcl --script hunch.lisp&

Yep!!, let’s checkout hunchentoot wheter it’s working

$ curl localhost:8080

If you see the hunchentoot’s response you’re ready to configure nginx to proxy web page requests to hunchentoot’s listening port. We’ll create a virtual host and proxy the requests that has paths end with ‘.html’ to the hunchentoot. First create a virtual server home directory.

$ cd
$ mkdir -p nginx/html/example.org/htdocs

Now open nginx configuration file and create a virtaul host configuration within the http node of the configuration file like below.

    server {
        listen       80;
        server_name  example.org;
        location / {
            root   html/example.org/htdocs;
        }
        rewrite ^(.*)/$ $1/index.html;
        location ~ .html$ {
            proxy_pass      http://127.0.0.1:8080;
        }
    }

You must now restart your nginx process to reload it’s configuration file

$ cd
$ sudo pkill nginx
$ cd nginx
$ sudo sbin/nginx

Yes! From now you have got a working lisp web server behind a powerful static content server. You should try it out by putting some static content to virtual host root directory and pointing your browser address to it. Please pay attention that we have proxied all html requests to hunchentoot, you should configure the nginx to handle any other regex.

Please if any command fails do not be upset cause i didn’t tried the commands line by line. I assume that you’re familiar unix/linux envirmenment and should correct if any instruction booms.

Hope this post help anyone interested…

public static void ReadFromConfig<T>(string key, ref T obj)
{
    string value = System.Configuration.ConfigurationManager.AppSettings[key];
    if (!string.IsNullOrEmpty(value)) obj = (T)System.Convert.ChangeType(value, typeof(T));
}

The usage is simple as itself.

// initialize a variable with a default value
int foo = 9;

// change the value of foo if any value assigned to it from config
// otherwise do nothing

ReadFromConfig<int>("foosetting", ref foo); 

Hope, this helps…

Since i decided to begin a C/C++ project, i thought that i must firstly prepare a development environment for my own convenience. So i began to dig about make tool for building an environment for myself. The most important feature i wish to implement was a proper directory structure. After hours reading GNU make manual, i found myself writing one that resolves header and source dependency between object files under a directory structure.

Here i share well parametrized and documented simplistic make file that supplies all my needs. The directory structure i use is:

./src
./src/greeter.cpp
./src/printer.cpp
./src/main.cpp
./bin
./makefile
./obj
./inc
./inc/printer.h
./inc/greeter.h

Since i’m not an make tool expert, please feel free to share if there is an absurdity or if you want extend it somehow

makefile:

# ------------------------------------------------------------------------------
# Copyright [2008] [Kadir PEKEL]
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# 	http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ------------------------------------------------------------------------------

# ------------------------------------------------------------------------------
# This simple GNU make file automates your build processes with simple
# dependency resolution under a DIRECTORY STRUCTURE.
# ------------------------------------------------------------------------------

# ------------------------------------------------------------------------------
# Name your own project
# ------------------------------------------------------------------------------

PROJECT = testProject

# ------------------------------------------------------------------------------
# Compiler parameters
# ------------------------------------------------------------------------------

CC = g++
CFLAGS = -Wall -g
LIBS = -lm

# ------------------------------------------------------------------------------
# Directory structure
# ------------------------------------------------------------------------------

SRC_DIR = src
OUT_DIR = bin
OBJ_DIR = obj
INC_DIR = inc

# ------------------------------------------------------------------------------
# Source file sufix
# ------------------------------------------------------------------------------

SUFFIX = .cpp

# ------------------------------------------------------------------------------
# You shouldn't modify below unless you do know what you are doing.
# -
# The thing that is going on here is that any source file under the source
# directory is compiled into its own object file under the objects directory.
# Dependency declaration of the object file to its corresponding source file
# is done automatically. Also, if your source file has got a header file with
# the same name in the include directory, it's also declared as a dependency
# for the related object file.
# ------------------------------------------------------------------------------

ALL_CFLAGS = -I. -I./$(INC_DIR)
OUTPUT = $(OUT_DIR)/$(PROJECT)

SRC_FILES = $(wildcard $(SRC_DIR)/*$(SUFFIX))

OBJ_FILES = $(patsubst $(SRC_DIR)/%$(SUFFIX), $(OBJ_DIR)/%.o, $(SRC_FILES))

INC_FILES = $(wildcard $(INC_DIR)/*.h)
INC_SRC_FILES = $(patsubst $(INC_DIR)/%.h, $(SRC_DIR)/%$(SUFFIX), $(INC_FILES))
INC_SRC_OBJ_FILES = $(patsubst $(SRC_DIR)/%$(SUFFIX), $(OBJ_DIR)/%.o, \
							$(INC_SRC_FILES))

_main: $(OBJ_FILES)
	$(CC) $(CFLAGS) $(ALL_CFLAGS) $(OBJ_FILES) -o $(OUTPUT) $(LIBS)

$(INC_SRC_OBJ_FILES): $(OBJ_DIR)/%.o: $(INC_DIR)/%.h

$(OBJ_DIR)/%.o: $(SRC_DIR)/%$(SUFFIX)
	$(CC) $(CFLAGS) $(ALL_CFLAGS) -c $< -o $@

# ------------------------------------------------------------------------------
# Default target that does the work when you type 'make' without pointing to
# any target
# ------------------------------------------------------------------------------

.DEFAULT : all
all: _main

# ------------------------------------------------------------------------------
# Put your cross dependiences here
# ------------------------------------------------------------------------------

$(OBJ_DIR)/greeter.o: $(INC_DIR)/printer.h
$(OBJ_DIR)/main.o: $(INC_DIR)/greeter.h

# ------------------------------------------------------------------------------
# Here you can write down custom targets and if you wish, you can make the
# default target dependent to yours. For ex: 'all: _main <your_target>'
# ------------------------------------------------------------------------------

.PHONY : clean
clean:
	-rm $(OBJ_DIR)/* $(OUT_DIR)/*

your_target:
	@echo "your target"

Disable mark occurrences

Open preferences (choose of project wide or global wide preferences is up to you) window and type “mark occurrences” in filter text box. Select the “Mark Occurrences” from the list box and remove the tick from “Mark occurrences of the selected element in the current file” check box.

Remove structured text validation

This time type “validation” into the filter text box in preferences window and select “Validation” from the list box. You’ ll see some of file types that are promised to be validated. Deselect validation ticks from all of the file types in the list (you can do it manually later if you want to). You’ll see a really very big difference in eclipse performance if you have big xml and wsdl files. For a example in my last J2EE project, my web.xml files contain 1400 and wsdl files contain thousands lines of text so eclipse couldn’t handle all the validations while computer memory is avarage

Do not use subclipse plug-in

Subclipse consumes so much system resources and effects eclipse performance greedily in big projects. If you could, consider not to use subclipse especially in projects that contain thousands of code kept in subversion source repository. It’s really become a very heavy-weight plug-in with heavy-weight code. You should feel better using subversion from the command line or from a seperate client

Consider converting your static code to a jar library

This advise can be possible more likely when you have static code automatically generated from static wsdl belongs to a web service. By this way you reduce the raw code size in project and use the code functionality from compiled classes to force eclipse to use fewer system resource.

Configure java virtual machine memory management start up arguments

In your eclipse.ini file, set -Xms40m and -Xmx256m args as your needs. This options define minimum and maximum memory usage bounds which passed to java virtual machine to manage eclipse application domain’s memory allocation tolerance. You can tweak this values and experiment your optimum eclipse speed.

Also if you have problems in eclipse’s memory management in Linux os environments like having lots of out of memory errors, you should define permgen space argument in eclipse.ini file. Setting this arguments as needed, you will have very few (or not) sudden memory exceptions. Try these -XX:PermSize=128m -XX:MaxPermSize=128m values if you got for about 1GB ram in your machine.

Finally the Eclipse version i currently have is 3.3.1.1.

htop

Htop is young and beautiful sister of top on gnu/linux. It views your processes on your system with resources they use in a more visual way that top does. Htop provides fully customizable pretty colored widgets like gauges that display total system resource usage.

Mutt

A superior e-mail, newsgroup client that works on terminal with a simple configuration and setup. If you love to work on terminal and want to access your incoming mails rapidly on terminal then you can easily adopt to Mutt

Finch

Finch is a terminal clone of famous IM client Pidgin and is developed again by Pidgin developers. It has almost the same visual apperance as Pidgin on X but this time on terminal. It uses ncurses libraries and worths to give it a try.

SnowNews

SnowNews is a simple feed reader that supports all versions of RSS and OPML. It uses libxml2 and ncurses libraries and has many smart features like cookie handling, auto http redirection, caching and colored customization.

Mpd - Ncmpc

Mpd is a music player daemon which publishes your music directory to a local port or to a shoutcast server. You can use it as a remote access point for your own music box on public or private network. It uses a special protocol implemeted by mpd developers. The client which i should suggest is ncmpc which plays your music box in your terminal enviroment with a simple nice looking interface. It has a sweet looking digital clock too.

Midnight Commander

Famous file manager for terminal environment which gets easy your file system operations, also you can login to ssh or ftp server to manage your remote files.

ccze

if you get blind in long logs, try them to colourize with ccze. It simply represents your logs in fancy colours to you. If you monitor your syslogs, apache logs or etc… so much, try this and don’t get lost in log hell.

BitchX

The most stable, commonly used, robust irc client in the world that sits on terminal. Also i want to share a quote from bitchX website :),

BitchX is the most popular IRC client among Unix systems. You can also use it on Windows, but if you had a Picasso painting, would you put it in the bathroom?

LoL…

screen

Probably you already have screen in your os if you use unix or unix like system and again probably you already know what screen is but i want to add it to this list for whom who doesn’t meet it if any. Screen is terminal multiplexer which means that you could have more than one interactive shell session in the same terminal screen. ıt’s very useful espacially when you open a remote shell session (ssh connection).

XMonad

It’s a fantastic X window manager for unix like systems which organizes windows in useful tiling algorithms. Windows will not float until you want and you’ll never miss a gnome - KDE like enviroment when you adopt this. I strongly suggest if you work in terminal emulator too much.

Finally,

So, i will add any application to this list if i remember or meet a new one. Feel free to share any if you think it must be in the list.

Update Dec 23, 2008 [Readers' contributions]

Extemp

Extemp is software, written for GNU/Linux, that uses the rss feeds of online newspapers to automatically download articles into a searchable database.

irssi

Irssi is a terminal based IRC client for UNIX systems. It also supports SILC and ICB protocols via plugins.

ncmpcpp

Ncmpcpp is almost exact clone of ncmpc but it contains some new features ncmpc doesn’t have. It’s been also rewritten from scratch in C++. Lack of some function (eg. tag editor) eventually pushed me to rewrite it as ncmpcpp and include these functions in it.

rtorrent

rtorrent is a BitTorrent client for ncurses, using the libtorrent library. The client and library is written in C++ with emphasis on speed and efficiency, while delivering equivalent features to those found in GUI based clients in an ncurses client.

tmux

tmux is a “terminal multiplexer”, it enables a number of terminals (or windows) to be accessed and controlled from a single terminal. tmux is intended to be a simple, modern, BSD-licensed alternative to programs such as GNU screen.

bitlbee

BitlBee brings IM (instant messaging) to IRC clients. It’s a great solution for people who have an IRC client running all the time and don’t want to run an additional MSN/AIM/whatever client.

sup

Sup is a console-based email client for people with a lot of email. It supports tagging, very fast full-text search, automatic contact- list management, custom code insertion via a hook system, and more.

weechat

WeeChat (Wee Enhanced Environment for Chat) is a free IRC client, fast and light, designed for many operating systems.

as a result of an accident all comments of it-box has lost, sorry for inconvenience

The big picture what i see is; there is a world of wonders which has a god (Microsoft), and prophets (Scott Guthrie, Scott Hanselman, Rob Connery : The last Prophet etc.) and we all the creatures. We are all crying out to our god:

God:

- Hoo Hoo hooo, give me your soul poor man, i will bless you. With your $acrifices, i would be appreciated.

Prayers:

- oh my god, please please give me food, i gave you my soul already, please give me some more.

Among all these my posts below in the blog, it’s interesting that why i’m talking about like that for now? Nowadays, everyone is aware of the ASP.NET MVC framework you know and all the Microsoft folks waiting for it in a lather (sadly me too!). Every other’s blog which i get in, get me out with the comments like

# re: .NET Web Product Roadmap (ASP.NET, Silverlight, IIS7)
Thursday, November 29, 2007 1:39 PM by Wallstreet

Give me silverlight 2.0 or give me death!

comment on Scott Guthrie’s “.NET Web Product Roadmap (ASP.NET, Silverlight, IIS7)” post

Kevin Isom
October 19th, 2007

But in saying all of that I can’t wait to get my hands on the MVC stuff as well.

comment on Rob Connery’s “Have We ASP.NET Geeks Lost Something?” post

These comments describe what i exactly trying to mention. I called this “followship”. We all ms guys playing around with the toys that Microsoft gave us. Everyone in the community looks for Scott Guthrie’s mouth for any lips movement. We are greedily waiting for MVC ctp deadline, when we eat it, our god will feed us with an another toy. What’s new here? MVC patterns exists for years and when Microsoft implements it it becomes one of holy grails. The followship buzzword sits on the right meaning for Microsoft community IMHO.

Then what is “fellowship” ? I won’t describe the the meaning what i load to, but i’ll gave you some steps to open your wings to gain full freedom and to open doors to no headed community.

• Get you pc Debian installed (or any other popular linux distro)
• Tell your distro’s packet manager (aptitude / apt-get) to install GCC, Python, Java and Eclipse IDE
• Tell your Eclipse IDE to install PyDev and CDT plug-ins
• Welcome to the fellowship of the free software

All you need to develop really real world software is on your hands for FREE even in enterprise solutions. Take it so:

1. Do you need speedy and useful developer tools on hand with a powerful shell? Welcome to *unix world!!!
2. Do you need speed? Take C/C++ programming language, learn it, experience it, tool it for your box, in Eclipse IDE with CDT plug-in.
3. Do you need efficiency? Take JAVA programming language, it won’t hurt you if you already code with C# also it has extensive Enterprise Frameworks like Spring, Struts and Eclipse IDE is the hometown for JAVA.
4. Do you need productivity? Take Python programming language, learn what it can save for your life time, again in famous Eclipse IDE with PyDev plug-in.

The all stack comes with real cross-platform comformance, they are not built for a specific platform like IIS or Windows api. They can extend on free Apache http server which works on all platforms, they can extend on Gtk+ GUI toolkit which works on all platforms etc. etc. etc. Give them a chance and with your individual skills you should be number one coder and developer with community and freedom support Believe in that here no one expects nothing from someone except some help without get paid.

And it’s time for you to face with your own realities…

Oh my, visual studio is really great IDE software, it makes all the things for me and now i could really dive into thinking the real problems of my software needs. But where is analytical thinking? Trying to solve the big picture puzzle of software is really an analytical thinking? I think the nature of computer programming is writing code that computers undertands and trying to solve problems by using innovational or existing algorithms. So the one who solves computional problems by algorithms is a computer programmer in a nutshell i think.

If you doubt, one main question have to be asked:

What is analytical thinking?

Here comes my second task for you that you have an array of digits those random positioned like this one:

int[] digits = new int[]{4, 5, 2, 7, 9, 3, 1, 8, 6};

So please give a try to sort this array ascending like this one {1, 2, 3, 4, 5, 6, 7, 8, 9 } without using Array.Sort() or any library method. This sounds very simple, is it? i can hear your laugh, saying how an easy thing this is. If you give a chance yourself to solve this problem i guess and suppose that you will find how much an analytical thinker you are.

*

Here i do not advocate that not to use visual studio or framework libraries, certainly they make our life more easier, but all programmers who says that i’m a programmer and have ability to solve any kind of problems by using my favorite programming language, must do training and exercise in basic algorithms, learn historical solutions to generic problematic issues and have a look at some framework base classes how things done.

I’m neither a perfect programmer nor a expert analytical thinker, but i’m trying to do and always give myself some time to understand some basic and base concepts. I guess and believe these will make me soon a good big picture puzzler, a senior developer or solution developer.

It could stand as an alternative for wsdl.exe tool of .NET Framework sdk or Visual Studio’s web referencing feature.

I won’t describe the code one by one, you can watch the poor and short video below and download the project source code at the and of this post to digg in somemore. It can cause reinspirations nicely, if you catch one please let me know, i would like to help. So enjoy it!

P.S: If you have interested in SubSonic project which is exactly popular nowadays, this work inspired a bit from it.

Download Project Source Code